Skip to main content

Home / AI readiness checklist

AI automation readiness checklist

Use this before requesting a proposal. Unresolved data, ownership, escalation, or measurement gaps should become explicit pilot risks.

Workflow

  • The task has a clear start and completion event
  • Volume and exceptions can be measured
  • A process owner can approve changes
  • The workflow is stable enough to pilot

Data

  • Required data sources are identified
  • Access and retention rules are documented
  • Representative test examples exist
  • Sensitive fields can be minimized or protected

Integration

  • Required systems expose a controlled interface
  • Actions can be logged and reversed where appropriate
  • Credentials can be scoped
  • Failures have a safe fallback

Risk

  • Human escalation conditions are defined
  • Incorrect output has a known impact
  • Consent requirements are understood
  • A person can stop the pilot quickly

Measurement

  • A pre-pilot baseline exists
  • Success and failure metrics are written
  • Cost includes human review and exceptions
  • A scale/no-scale date is agreed

How to use the result

Choose the workflow with the clearest owner, safest failure mode, best baseline, and smallest integration surface—not necessarily the largest theoretical saving.

Discuss a contained pilot